Diagnosing Dysautonomia After Acute Traumatic Brain Injury: Evidence for Overresponsiveness to Afferent Stimuli

Article Outline

• Abstract
• Methods
  • Sample Characteristics
  • Procedure
  • HRV Analysis
  • Event-Related Heart Rate
  • Analysis
• Results
  • Group Demographic and Resting Cardiac Data
  • Stimulus-Dependent Data
• Discussion
  • Study Limitations
• Conclusions
• References
• Copyright

Abstract 

Baguley IJ, Nott MT, Slewa-Younan S, Heriseanu RE, Perkes IE. Diagnosing dysautonomia after acute traumatic brain injury: evidence for overresponsiveness to afferent stimuli.

Objective

To differentiate between traumatic brain injury (TBI) subjects with normal and elevated autonomic activity by quantifying cardiac responsivity to nociceptive stimuli and to determine the utility of heart rate variability (HRV) and event-related heart rate changes in diagnosing dysautonomia.

Design

Prospective cohort study.

Setting

Intensive care unit in a tertiary metropolitan trauma center.

Participants

Adults (N=27) with TBI recruited from 79 consecutive TBI admissions comprising 16 autonomically aroused and 11 control subjects matched by age, sex, and injury severity.

Interventions

None.

Main Outcome Measures

Immediate: pattern of autonomic changes indexed by HRV and event-related heart rate after nociceptive stimuli. Six months: length of stay, Glasgow Coma Scale, and Disability Rating Scale.

Results

Heart rate changes (for both HRV and event-related heart rate) were associated with the diagnostic group and 6-month outcome when evaluated pre- and poststimulus but not when evaluated at rest. When assessed on day 7 postinjury, the comparison of HRV and heart rate parameters suggested an overresponsivity to nociceptive stimuli in dysautonomic subjects. These subjects showed a 2-fold increase in mean heart rate relative to subjects with sympathetic arousal of short duration (16% vs 8%), and a 6-fold increase over nonaroused control subjects. Data suggest that post-TBI sympathetic arousal is a spectrum disorder comprising, at one end, a short-duration syndrome and, at the other end, a dramatic, severe sympathetic and motor overactivity syndrome that continued for many months postinjury and associated with a significantly worse 6-month outcome. These findings suggest that it is not the presence of reactivity per se but rather the failure of processes to control for overreactivity that contributes to dysautonomic storming.

Conclusions

This study provides empirical evidence that dysautonomic subjects show overresponsiveness to afferent stimuli. Findings from this study suggest an evidence-driven revision of diagnostic criteria and a simple clinical algorithm for the improved identification of cases.

Key Words: Autonomic nervous system, Brain injuries, Diagnosis, Heart rate, Rehabilitation

List of Abbreviations: bpm, beats per minute, DRS, Disability Rating Scale, EKG, electrocardiogram, ETT, endotracheal tube, GCS, Glasgow Coma Scale, HF, high frequency, HRV, heart rate variability, ICU, intensive care unit, LF, low frequency, PAID, paroxysmal autonomic instability with dystonia, RR, interbeat intervals, TBI, traumatic brain injury, WCC, white cell count

ACUTE ACQUIRED BRAIN INJURY commonly causes short-term elevations of autonomic parameters in the ICU. A subgroup of these patients go on to develop a syndrome of persistent paroxysmal sympathetic and motor overactivity that is known by various names including dysautonomia, sympathetic storms, and PAID.¹ Although dysautonomia can result from any form of acquired brain injury, it occurs most frequently after moderate to severe TBI,¹ making this an accessible model for investigating the phenomenon.

After TBI, 62% to 92% of patients exhibit at least 1 significantly increased parameter such as heart rate of 120 bpm or higher, respiratory rate 30 breaths/minute or more, or systolic blood pressure of 140mmHg or higher.², ³ In approximately one quarter to one third of cases,², ⁴ these increases occur simultaneously in a manner suggesting a paroxysmal and widespread increase in sympathetic activation. The onset of such paroxysmal episodes has been reported to coincide with the withdrawal of sedation², ⁵ and, for most subjects, represents a self-limiting disorder of uncertain significance.

However, in 8% to 12% of patients,², ⁶, ⁷ these changes result in persistent sympathetic and motor overactivity, fulfilling the diagnostic criteria for dysautonomia.⁵ These dysautonomic patients have significantly poorer outcomes,⁵ greater health care costs,² and are at risk of increased morbidity if not recognized and treated appropriately.⁵, ⁸, ⁹ However, at the present time, the identification of dysautonomia is reliant on clinical observation of the specific symptom cluster, and the diagnosis remains one of exclusion.¹, ¹⁰, ¹¹

An association between the onset of paroxysmal overactivity and afferent stimuli has recently been highlighted in dysautonomic patients.¹² Stimuli reported to induce paroxysms may include pain; endotracheal suctioning; passive movement such as turning, bathing, and muscle stretching,⁵, ¹⁰, ¹³, ¹⁴, ¹⁵, ¹⁶, ¹⁷, ¹⁸, ¹⁹, ²⁰, ²¹ constipation or urinary retention,¹⁴, ²² and environmental/emotional stimuli.²³, ²⁴ Overall, these observations suggest that paroxysms can be driven by exogenous stimuli, not all of which would be considered nociceptive in the usual sense. It has consequently been hypothesized that the abnormal processing of such afferent stimuli is the “sine qua non” of dysautonomia.¹², ²⁰

Recently, efforts have been made to advance the diagnostic process from anecdotal clinical observation to quantitative physiologic measurement. Of particular relevance, paroxysm-related tachycardia appears to be the most persistent autonomic abnormality observed in dysautonomia.², ⁵ Abnormal cardiac control mechanisms (assessed via HRV paradigms) have been found to differentiate dysautonomic from nondysautonomic subjects and healthy controls at 3 months postinjury, changes that persist for at least 14 months postinjury.⁹ In other research in acute TBI, reduced HRV has correlated with greater injury severity,²⁵, ²⁶, ²⁷ poorer global outcome,²⁸ and death,²⁹ with improving HRV parameters being linked to improving neurologic function.³⁰ However, in contrast, some studies⁶, ³¹ have not found an association between autonomic dysfunction and poorer functional outcome. To date, reasons for such divergent evidence have not been elucidated.

Drawing from this background, it was hypothesized that longitudinal cardiac monitoring would provide a means of investigating sympathetic overactivity after standardized afferent stimuli and provide insights into the pathophysiology of dysautonomia. Specifically, this study sought (1) to determine how HRV and outcome parameters differ between subjects with normal and elevated autonomic responses, (2) to determine whether event-related heart rate and HRV changes could provide diagnostic tools for identifying dysautonomia, and (3) to evaluate the hypothesis that dysautonomic subjects exhibit a disproportionate paroxysmal response to nociceptive stimuli.

Back to Article Outline

Methods 

Sample Characteristics 

This study used observational data on a subsample of 27 subjects drawn from a larger prospective cohort study involving 79 subjects with moderate to severe TBI recruited from consecutive ICU admissions to a tertiary-level hospital over a 2-year period. Sample characteristics and recruitment procedures are detailed elsewhere.² Subjects were classified into autonomically aroused (n=19) or standard TBI (n=60) based on the presence or absence of elevated physiologic parameters (eg, heart rate >120 bpm, respiratory rate >30 breaths/min, temperature >39°C, and blood pressure above 160mmHg) on day 7 postinjury. The subset of 27 subjects in this study comprised the available autonomically aroused subjects (n=16, stimulus-dependent EKG data were not available for 3 subjects from the original autonomically aroused group) and a matched sample of standard TBI subjects from the larger sample (n=11, matched on age, sex, and best GCS score in the first 24 hours). Subjects with diseases likely to modify autonomic control of heart rate (eg, diabetes or other autonomic neuropathy) were excluded.

Differentiation between patients with self-limiting autonomic arousal and the prolonged variant of dysautonomia occurred on day 14. Six subjects in the autonomically aroused group continued to show autonomic arousal symptoms comprising simultaneous, paroxysmal increases in at least 5 of 7 clinical features at least daily: heart rate, respiratory rate, temperature, blood pressure, dystonia, posturing, and sweating,⁵ and were consequently classified as dysautonomic (n=6). Autonomic hyperactivity had ceased in the remaining 10 autonomically aroused subjects at day 14; therefore, these patients were classified as nondysautonomic (n=10).

Procedure 

After approval of the project from the local institutional ethics committee, informed consent was gained from each subject's next of kin. Demographic and injury-related data (eg, age, sex, GCS, DRS³²) were collected from the medical record.

Previous research⁵, ⁹ has shown that heart rate changes are the most consistent and prolonged autonomic change seen in dysautonomic subjects.⁵, ⁹ As a consequence, heart rate response to stimuli was selected as being the most sensitive marker of sympathetic autonomic dyscontrol for this study. Correspondingly, on day 7 postinjury, continuous EKG data were collected at 1000Hz for approximately 6 minutes before and after the application of a routine noxious stimulus (ETT suctioning [n=26] or mouth care [n=1]) with subjects lying supine via an ADInstruments Powerlab system.^a ETT suctioning was completed by nursing staff as a component of routine care. Subjects were not actively storming at the time of stimulus application. For each subject, EKG data were then processed offline by using 2 analysis techniques: spectral HRV and an event-related heart rate analysis (described later). Details of each subject's medications, WCC, GCS, and DRS were also collected from the medical record on day 7 postinjury.

On day 14, subjects were classified into dysautonomic and nondysautonomic groups as per procedures outlined earlier. Stimulus response data were therefore collected before group allocation, minimizing bias in data collection. Finally, outcome variables were collected at 6 months postinjury including posttraumatic amnesia duration, DRS, GCS scores, and ICU/total inpatient length of stay. To minimize observer bias, outcome variables were collected by clinical staff blinded to subject classification.

HRV Analysis 

Linear HRV analyses (frequency and spectral analysis) were performed by using the Heart Rate Variability module of ADInstruments Chart (V5.0) software.^a Spectral analysis was performed via Fast Fourier Transform of resampled data via a Welch window with 50% overlap.⁹ Spectral power data were processed for sympathetically mediated LF (range, .04–.15Hz) and parasympathetic HF power (range, .15–.4Hz). LF and HF data were normalized to remove the effect of different time-collection intervals and improve data portability between differing HRV analysis techniques.³³ From these data, the LF/HF ratio (representing the ratio of sympathetic and parasympathetic influences on the heart) was derived for 5-minute intervals immediately before and after stimulus.³⁴ The 5-minute interval was selected as a compromise between collecting meaningful HRV data and limiting the cardiac response measurement to the transitory period evoked by the stimulus. Thus, a minimum period was expected to limit the influence of other environmental stimuli that could not be controlled in this observational study.

Event-Related Heart Rate 

The event-related heart rate analysis used sequential RR derived from the entire period of observation. RR data (in ms) was referenced against the stimulus (labeled as beat zero). A preliminary analysis showed that 25 of the 27 subjects produced an initial increase in heart rate after the stimulus. The primary maximum heart rate response occurred a mean ± SD of 32.6±9.7 beats poststimulus, with no significant variation in the timing of the peak between groups (F=1.11, P=.35).

On this basis, and in contrast to the HRV analysis, it was decided to focus the event-related heart rate analysis on the 100-beat intervals before and after the stimulus. The mean heart rate was subsequently calculated for the 100 beats pre- and 100 beats poststimulus (termed pre/poststimulus 100 beats) for each subject. To control for intersubject variability in baseline heart rate, each subject's data were normalized against his/her mean RR interval for the 100 prestimulus beats. Group mean data for this analysis is displayed as percentage change around a mean of 0%.

Analysis 

Between-group differences were assessed for the standard TBI and autonomically aroused groups on demographic, injury, resting cardiac, and outcome variables. Independent t tests or Mann-Whitney U tests were performed on normally and nonnormally distributed data, respectively. Between-group differences were examined in the dysautonomic and nondysautonomic groups for outcome variables using Mann-Whitney U tests.

Repeated-measures analysis of variances was used to determine between-group differences as a function of time. First, the standard TBI group was compared with the autonomically aroused group on cardiac parameters. Second, the dysautonomic and nondysautonomic subgroups were compared on their pre- and poststimulus responses. Cardiac data (HRV total, LF and HF power, and the LF/HF ratio) were naturally log transformed to achieve a normal distribution. The mean heart rate data during the 5-minute interval pre- and poststimulus and the mean heart rate for the 100 beats pre/poststimulus were normally distributed and thus not transformed. Finally, the association between the LF/HF ratio and heart rate was examined (protocol per Baguley et al⁹) to investigate between-group differences in uncoupling of these parameters.

Calculations were made using the SPSS package (version 15.0).^b Results were considered significant when P was less than or equal to .05 for all analyses. The effect size was measured by using partial eta squared³⁵ with values of (partial η²=.09) considered a medium effect size and values of (partial η²=.25) a large effect size.³⁶

Back to Article Outline

Results 

Group Demographic and Resting Cardiac Data 

There were no significant differences between the matched standard TBI and autonomically aroused groups in terms of demographic, injury, and resting cardiac variables (table 1). Similarly, between-group differences were not significant for any outcome variables. Mean ± SD WCC for the entire sample (n=27) was 12.81±5.1. There was no significant difference (t=1.15, P=.26) between mean WCC for the standard TBI (n=11; mean ± SD, 14.15±1.8) and autonomically aroused groups (n=16; mean ± SD, 11.88±4.4).

Table 1. Group Demographic, Injury, Resting Cardiac, and Outcome Variables

Variables	sTBI (n=11)	AA (n=16)
Demographic variables
Age (y)	27.9±12.9	30.3±13.7
Sex
Male, n (%)	10(91)	15(94)
Female, n (%)	1(9)	1(6)
Injury severity variables
GCS (best <24h)	5.6±3.5	5.1±2.8
GCS (day 7)	6.0±3.8	7.8±2.0
DRS (best <24h)	25.1±6.1	26.9±2.4
DRS (day 7)	24.9±5.2	24.3±2.0
HRV variables at rest
HRV total power	1056.6±2301.5	1513.8±2821.5
LF (ms2)	182.0±370.3	274.0±555.1
HF (ms2)	402.3±1193.1	440.6±1538.7
LF normalized (ms2)	49.4±25.7	61.9±25.3
HF normalized (ms2)	50.6±25.7	38.1±25.3
LF/HF ratio	1.5±1.3	3.7±4.0
Mean HR (bpm)	82.9±15.9	89.0±14.6
Outcome variables
ICU LOS (d)	15.3±3.4	15.9±6.9
PTA duration (d)	42.1±27.2	37.5±13.7
Total LOS (d)	159.8±148.7	154.7±222.7
GCS (6mo)	14.6±0.7	13.7±2.5
DRS (6mo)	7.2±7.7	7.3±9.6

NOTE. Values are mean ± SD, unless otherwise noted.

Abbreviations: AA, autonomically aroused; HR, heart rate; LOS, length of stay; PTA, posttraumatic amnesia; sTBI, standard TBI.

Subjects in the standard TBI group were more frequently receiving morphine, midazolam, and noradrenaline (table 2). Further statistical analysis was prohibited by small cell sizes; however, clinical review of the data suggested that subjects receiving continuous morphine and midazolam infusions tended to show minimal response to stimulus. The use of fentanyl, propofol, β-blockers, and clonidine did not differ significantly between groups.

Table 2. Medications in Use at Day 7 by Subjects in Each Group

Medications	sTBI (n=11)	AA (n=16)
Morphine	8(73)	3(19)
Midazolam	8(73)	3(19)
Fentanyl	1(9)	3(19)
Propofol	10(91)	12(75)
Betablocker	1(9)	5(31)
Clonidine	3(27)	9(56)
Noradrenaline	6(55)	1(6)

NOTE. Values are n (%). Use of autonomically active drugs and sedation/paralysis medication by the standard TBI group and autonomically aroused group.

Abbreviations: AA, autonomically aroused group; sTBI, standard TBI group.

Stimulus-Dependent Data 

A dual-step analysis evaluated the hypothesis that dysautonomic subjects exhibit a disproportionate paroxysmal response to nociceptive stimuli. First, pre-/poststimulus differences between the standard TBI group and the autonomically aroused group were examined (table 3). Applying a noxious stimulus produced significant group and time interactions between the standard TBI and autonomically aroused groups for HF power (F=4.8, P=.038), LF/HF ratio (F=9.1, P=.006), and the mean heart rate of the pre-/poststimulus of 100 beats (F=8.7, P=.007). Poststimulus HF power in the autonomically aroused group decreased significantly, contributing to a significant increase in LF/HF ratio. The mean heart rate (100-beat interval) of the autonomically aroused group increased significantly after noxious stimulus, rising by a mean of 9.1 bpm from the prestimulus phase as compared with 2.1 bpm for the standard TBI group.

Table 3. Pre- and Poststimulus Measures for Standard TBI and Autonomically Aroused Groups

NOTE. Values are mean ± SD. Repeated measures analysis of variance (with effect size) comparison between the standard TBI group and autonomically aroused group, pre- and poststimulus.

Abbreviations: AA, autonomically aroused group; Ln_HF, natural log transform of HRV high frequency power; Ln_LF, natural log transform of HRV low frequency power; Ln_LF/HF ratio, natural log transform of LF/HF ratio; Ln_HRV total power, natural log transform of HRV total power; sTBI, standard TBI group,

The autonomically aroused group was then divided into nondysautonomic and dysautonomic to test the hypothesis that dysautonomic subjects exhibit a disproportionate afferent stimulus response. Only interactions that were statistically significant in the previous analysis were included in the second analysis (table 4). A statistically significant difference was found for the change in the mean heart rate during the pre-/poststimulus 100-beat period (F=4.9, P=.044) between the nondysautonomic and dysautonomic groups. This equated to a mean 6.3-bpm increase for the nondysautonomic group compared with 13.8 bpm in the dysautonomic group over the 100 poststimulus beats. This produced a large between-group difference as measured by effect size (partial η²=.29). In contrast to the pre-/poststimulus 100-beat data, stimulus effects between the nondysautonomic and dysautonomic groups for the remaining cardiac variables were not statistically significant.

Table 4. Pre- and Poststimulus Measures for Nondysautonomic and Dysautonomic Groups

NOTE. Values are mean ± SD. Repeated measures analysis of variance (with effect size) comparison between nondysautonomic group and dysautonomic group, pre- and poststimulus.

Abbreviations: DYS−, nondysautonomic group; DYS+, dysautonomic group; HR, heart rate; Ln_HF, natural logarithmic transformation of high frequency power; Ln_LF/HF, the low frequency/high frequency power ratio.

Figure 1 shows the relationship between group mean LF/HF ratio and mean 100-beat heart rate data for each group pre- and poststimulus. Consistent with the data reported previously, the standard TBI group showed little change in either variable before and after the stimulus. Prestimulus, both parameters were elevated for the autonomically aroused subgroups to a similar degree (LF/HF=5.2, heart rate=89 bpm for the dysautonomic group compared with LF/HF=2.8, heart rate=90 bpm for nondysautonomic). Poststimulus, the dysautonomic group showed a 2-fold increase in both variables compared with the nondysautonomic group (dysautonomic LF/HF=9.5, heart rate=103 bpm vs nondysautonomic LF/HF=5.2, heart rate=97 bpm).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Plot of group mean LF/HF ratio and heart rate for 100 beats pre/poststimulus. NOTE. Displays the relationship between LF/HF ratio and heart rate for each group: sTBI group, DYS− group, DYS+ group, before and after stimulus application. The sTBI group showed minimal change. DYS− and DYS+ groups showed collinear response patterns, with the DYS+ group achieving a 2-fold increase over the DYS− group. Abbreviations: DYS+, dysautonomic; DYS−, nondysautonomic; sTBI, standard TBI.

Normalized data for the 100-beat pre-/poststimulus interval for each group (standard TBI, nondysautonomic, dysautonomic) is presented in figure 2. The standard TBI group's heart rate response pattern showed a mean 2% increase poststimulus, predominantly from a transitory heart rate rise. The nondysautonomic group showed an 8% increase in the mean heart rate with a similar transitory poststimulus response pattern. The nondysautonomic group's response then diverged from the standard TBI group, showing a late tendency to increase. In contrast, the dysautonomic group showed a rapid and sustained 16% increase in the mean heart rate after the stimulus; it remained at this elevated level for 100 beats.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  Normalized group beat-to-beat change in heart rate for 100 beats pre-/poststimulus. The figure highlights the differences in immediate HR response to the stimulus across each group. Poststimulus HR increased by 2% in the sTBI group, 8% in the DYS− group, and by 16% in the DYS+ group. Abbreviations: DYS+, dysautonomic; DYS−, nondysautonomic; sTBI, standard TBI.

Finally, outcome data for the dysautonomic and nondysautonomic subgroups were compared to determine if significant differences existed between the groups (table 5). Two subjects in the dysautonomic group did not have late outcome data (deceased), and their data were excluded from analysis of these variables. Outcome variables were significantly worse for the dysautonomic group, who achieved lower GCS (z=–3.23, P=.001) and higher DRS scores at 6 months postinjury (z=–2.76, P= .004) and remained in hospital on average 8 times longer (z=–2.83, P=.005).

Table 5. Outcomes Achieved by Nondysautonomic and Dysautonomic Groups

NOTE. Values are mean ± SD. Mann Whitney U test comparisons for length of stay.

Abbreviations: DYS−, nondysautonomic group; DYS+, dysautonomic group.

Back to Article Outline

Discussion 

This study evaluated the effect of a semistandardized nociceptive stimulus on physiologic variables in moderate/severe TBI survivors. The 16 subjects in the autonomically aroused group were not actively storming when the stimulus was applied. These subjects were subclassified as either dysautonomic or nondysautonomic based on clinical evidence of simultaneous, paroxysmal sympathetic overactivity on day 14 postinjury. Turning to the hypotheses of this article, the first goal was to determine how HRV and outcome parameters differed between subjects based on their autonomic responses. On day 7 postinjury and examined at rest, there were no significant differences in cardiac parameters between the autonomically aroused and standard TBI groups. However, analysis of cardiac data pre- and poststimulus revealed significant group by time differences. ETT suctioning produced a significant reduction in HF power and increased LF/HF ratio in the autonomically aroused group relative to the standard TBI group. Furthermore, mean heart rate increases observed during the 100-beat intervals pre-/poststimulus were significantly greater in the autonomically aroused group relative to the standard TBI group. These changes were not associated with significant WCC differences, suggesting that infection did not explain group differences. However, it is worth noting that dysautonomic subjects may have concomitant acute medical conditions that need to be identified and treated.¹⁸

To further examine these poststimulus cardiac changes, dysautonomic subjects were compared with nondysautonomic subjects. Here, the dysautonomic subgroup showed a different pattern of response to the stimulus, resulting in a significantly greater increase in pre-/poststimulus heart rate (averaged over the 100-beat interval) compared with the nondysautonomic group. This increase equated to a 2-fold increase in the mean heart rate for dysautonomic relative to the nondysautonomic subjects (16% vs 8%, respectively). Thus, although pre- and poststimulus data revealed significant between-group differences, evaluation of these same variables at rest did not.

Previous research³⁷, ³⁸ has also noted increased heart rate and blood pressure (markers of sympathetic arousal) after ETT suctioning in survivors of TBI in the ICU. The current research extends this finding by showing that sympathetic responsiveness is moderated by dysautonomic diagnostic status. Taken together, these results suggest that sympathetic reactivity is a normal response to stimulus in the early postinjury period. Previous research¹⁸ suggested that the lack of an observed response in the standard TBI group is most likely because of the sedation regimen used in this group. The different patterns and greater degree of reactivity in dysautonomic subjects is in keeping with suggestions that dysautonomia represents an overreactivity to the stimulus.³⁹ This leads to speculation that it is not the presence of reactivity per se but rather the failure of processes to regain control over this reactivity that may produce sympathetic storming in the dysautonomic group.

The second goal of this study was to determine the possible diagnostic utility of event-related heart rate and HRV changes for identifying dysautonomia. The significant pre-post stimulus differences observed between the dysautonomic and nondysautonomic groups provide preliminary support for the second hypothesis at a group level. Furthermore, data in this study identify a number of causes of between-subject heterogeneity that could explain discrepancies in previous research results. For example, minor variations in the timing of HRV data collection relative to time postinjury, recency of prior stimulation, the depth and pharmacologic nature of sedation, variability in cardioactive medication use, and dysautonomia status could all modify a subject's response pattern. Minimizing heterogeneity can then be achieved by evaluating poststimulus responses against prestimulus baseline data, where each subject acts as his/her own control. Quantitating the role that nociceptive stimuli play in provoking dysautonomic paroxysms also represents a refinement of previous investigatory protocols.⁹ In previous research, diagnostic thresholds for dysautonomia have been selected in an ad hoc fashion,¹ and it has been argued that fixed physiologic values (such as those required to diagnose PAID⁴⁰) introduce unnecessary difficulties.¹, ⁸, ⁴¹ Data from the current study suggest that assessing for paroxysmal, simultaneous sympathetic and motor overreactivity in response to nociceptive stimuli rather than using fixed threshold values may provide a more appropriate vehicle to identify those at risk of developing dysautonomia in subjects who are not actively storming at the time of assessment. Furthermore, the study suggests that cardiac parameters may be a simple mechanism to evaluate sympathetic overreactivity in this patient group.

The final goal of this study was to evaluate the hypothesis that dysautonomic subjects exhibit a disproportionate increase in heart rate in response to nociceptive stimuli. Evidence for this hypothesis is seen in figure 1 in which the dysautonomic group showed a 2-fold increase in mean heart rate versus LF/HF ratio relative to the nondysautonomic group and a 6-fold increase over standard TBI group from equivalent stimuli. This finding corroborates earlier HRV research by our group in dysautonomic subjects 11 weeks postinjury.⁹ However, the colinearity of dysautonomic and nondysautonomic response patterns in figure 1 suggests a physiologic overresponsiveness to afferent stimuli rather than an uncoupling of LF/HF ratio and heart rate as hypothesized previously.⁹ The overresponsiveness and the confirmation of triggering provide the first quantitative support for the central hypothesis of the Excitatory:Inhibitory Ratio Model of dysautonomia.³⁹

Findings from this study suggest an evidence-driven revision of diagnostic criteria and a simple clinical algorithm for improved identification of cases. Specifically, the data suggest that at-risk subjects would be those who exhibit a persistent overresponsiveness to nociceptive stimuli early postinjury. A diagnosis of dysautonomia would then follow if the subject continued to show simultaneous, paroxysmal increases in 5 of the 7 motor and sympathetic features linked to dysautonomia⁵ with stimulus overreactivity for 2 or more weeks postinjury. Accepting this suggestion requires further cross-validation studies in future research, which should also assess the sensitivity and specificity of this method.

Study Limitations 

As with virtually all dysautonomia research to date, the primary limitation of this study is the relatively small sample size. However, in the context of dysautonomia research, the parent study represents 1 of only 2 prospective studies into the condition with reasonable sample size.², ⁴ This limitation can only be overcome through multicenter research, which has been hampered by confused nomenclature and anecdotally derived diagnostic criteria.¹, ², ⁵ Another limitation of this study was the a priori decision to use a fixed time point (day 7 postinjury) to measure cardiac variables rather than selecting the withdrawal of sedation. By adopting this fixed time point, some subjects may have been defined as standard TBI only to become autonomically aroused after sedation withdrawal. However, the techniques reported herewith represent a novel mechanism to minimize sample heterogeneity and improve recognition of dysautonomic subjects in future research.

Back to Article Outline

Conclusions 

This article presents evidence to suggest that dysautonomia is a spectrum disorder comprising, at one end, a short duration syndrome that does not appear to have significant negative consequences for recovery. At the other end of the spectrum, dysautonomia presents with dramatic, severe sympathetic and motor overactivity that can continue for many months postinjury and is associated with poor outcome. Long-duration dysautonomia appears to be identifiable via the heart rate response pattern to ETT suctioning on day 7 postinjury, despite the presence or absence of regular sedation.

This study extends previous research by suggesting that the day 7 response pattern measured by event-related heart rate and HRV differentiates subjects with and without long duration dysautonomia. This article also presents the first quantitative data of an association between dysautonomia and an overreactivity to afferent stimuli. By using these data, it is proposed that the clinical diagnostic criteria for dysautonomia be revised to incorporate evidence of persistent overreactivity to afferent stimuli as a predisposing factor to developing clinical dysautonomia. The relative infrequency of dysautonomia suggests that consistent diagnostic criteria and multicenter studies should be used in future research.

Suppliers

Back to Article Outline

References 

1. Baguley IJ. Autonomic complications following central nervous system injury. Semin Neurol. 2008;28:716–725
   • View In Article
   • CrossRef
2. Baguley IJ, Slewa-Younan S, Heriseanu RE, Nott MT, Mudaliar Y, Nayyar V. The incidence of dysautonomia and its relationship with autonomic arousal following traumatic brain injury. Brain Inj. 2007;21:1175–1182
   • View In Article
   • CrossRef
3. Jennett B, Teasdale G. Management of head injuries. Philadelphia: F.A. Davis Co; 1981;p 84-326
   • View In Article
4. Rabinstein AA. Paroxysmal sympathetic hyperactivity in the neurological intensive care unit. Neurol Res. 2008;29:680–682
   • View In Article
   • CrossRef
5. Baguley IJ, Nicholls JL, Felmingham KL, Crooks J, Gurka JA, Wade LD. Dysautonomia after traumatic brain injury: a forgotten syndrome?. J Neurol Neurosurg Psychiatry. 1999;67:39–43
   • View In Article
   • MEDLINE
   • CrossRef
6. Fernández-Ortega JF, Prieto-Palomino MA, Muñoz-López A, Lebron-Gallardo M, Cabrera-Ortiz H, Quesada-García G. Prognostic influence and computed tomography findings in dysautonomic crises after traumatic brain injury. J Trauma. 2006;61:1129–1133
   • View In Article
   • MEDLINE
7. Hendricks HT, Guerts ACH, van Ginnekin BC, Heeren AJ, Vos PE. Brain injury severity and autonomic dysregulation accurately predict heterotopic ossification in patients with traumatic brain injury. Clin Rehab. 2007;21:545–553
   • View In Article
8. Rabinstein AA. Paroxysmal autonomic instability after brain injury. Arch Neurol. 2004;61:1625
   • View In Article
   • MEDLINE
   • CrossRef
9. Baguley IJ, Heriseanu RE, Felmingham KL, Cameron ID. Dysautonomia and heart rate variability following severe traumatic brain injury. Brain Inj. 2006;20:437–444
   • View In Article
   • MEDLINE
   • CrossRef
10. Lemke DM. Riding out the storm: sympathetic storming after traumatic brain injury. J Neurosc Nurs. 2004;36:4–9
    • View In Article
11. Oh SJ, Hong YK, Song E. Paroxysmal autonomic dysregulation with fever that was controlled by propanolol in a brain neoplasm patient. Korean J Int Med. 2007;22:51–54
    • View In Article
12. Baguley IJ, Heriseanu RE, Cameron ID, Nott MT, Slewa-Younan S. A critical review of the pathophysiology of dysautonomia following traumatic brain injury. Neurocrit Care. 2008;8:293–300
    • View In Article
    • CrossRef
13. Penfield W, Jasper H. Autonomic seizures (Epilepsy and the functional anatomy of the human brain). In: London: J & A Churchill Ltd; 1954;p. 412–437
    • View In Article
14. Sandel ME, Abrams PL, Horn LJ. Hypertension after brain injury: case report. Arch Phys Med Rehabil. 1986;67:469–472
    • View In Article
    • MEDLINE
15. Boeve BF, Wijdicks EF, Benarroch EE, Schmidt KD. Paroxysmal sympathetic storms (“diencephalic seizures”) after severe diffuse axonal head injury. Mayo Clin Proc. 1998;73:148–152
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (5067 KB)
    • CrossRef
16. Russo RN, O'Flaherty S. Bromocriptine for the management of autonomic dysfunction after severe traumatic brain injury. J Paediatr Child Health. 2000;36:283–285
    • View In Article
    • MEDLINE
    • CrossRef
17. Cuny E, Richer E, Castel JP. Dysautonomia syndrome in the acute recovery phase after traumatic brain injury: relief with intrathecal baclofen therapy. Brain Inj. 2001;15:917–925
    • View In Article
    • MEDLINE
    • CrossRef
18. Baguley IJ, Cameron ID, Green AM, Slewa-Younan S, Marosszeky JE, Gurka JA. Pharmacological management of dysautonomia following traumatic brain injury. Brain Inj. 2004;18:409–417
    • View In Article
    • MEDLINE
    • CrossRef
19. Diesing TS, Wijdicks EF. Arc de cercle and dysautonomia from anoxic injury. Mov Disord. 2006;21:868–869
    • View In Article
    • MEDLINE
    • CrossRef
20. Baguley IJ, Heriseanu RE, Gurka JA, Nordenbo A, Cameron ID. Gabapentin in the management of dysautonomia following severe traumatic brain injury: a case series. J Neurology Neurosurg Psychiatry. 2007;78:539–541
    • View In Article
21. Tong C, Konig MW, Roberts PR, Tatter SB, Li X. Autonomic dysfunction secondary to intracerebral haemorrhage. Anesth Analg. 2000;91:1450–1451
    • View In Article
    • MEDLINE
    • CrossRef
22. Srinivasan S, Lim CC, Thirugnanam U. Paroxysmal autonomic instability with dystonia. Clin Auton Res. 2007;17:378–381
    • View In Article
    • CrossRef
23. Bhigjee AI, Ames FR, Rutherford GS. Adult aqueduct stenosis and diencephalic epilepsy (A case report). J Neurol Sci. 1985;71:77–89
    • View In Article
    • MEDLINE
    • CrossRef
24. Lemke DM. Sympathetic storming after severe traumatic brain injury. Crit Care Nurse. 2007;27:30–37
    • View In Article
    • MEDLINE
25. Biswas AK, Scott WA, Sommerauer JF, Luckett PM. Heart rate variability after acute traumatic brain injury in children. Crit Care Med. 2000;28:3907–3912
    • View In Article
    • MEDLINE
    • CrossRef
26. Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the autonomic and cardiovascular systems in acute brain injury. Am J Physiol. 1998;275:R1287–R1292
    • View In Article
    • MEDLINE
27. Rapenne T, Moreau D, Lenfant F, Boggio V, Cottin Y, Freysz M. Could heart rate variability analysis become an early predictor of imminent brain death? (A pilot study). Anesth Analg. 2000;91:329–336
    • View In Article
    • MEDLINE
    • CrossRef
28. Winchell RJ, Hoyt DB. Analysis of heart-rate variability: a noninvasive predictor of death and poor outcome in patients with severe head injury. J Trauma. 1997;43:927–933
    • View In Article
    • MEDLINE
29. Norris PR, Ozdas A, Cao H, et al. Cardiac uncoupling and heart rate variability stratify ICU patients by mortality: a study of 2088 trauma patients. Ann Surg. 2006;243:804–812
    • View In Article
    • MEDLINE
    • CrossRef
30. Wijnen V, Heutink M, van Boxtel G, Eilander H, de Gelder B. Autonomic reactivity to sensory stimulation is related to consciousness level after severe traumatic brain injury. Clin Neurophysiol. 2006;117:1794–1807
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (352 KB)
    • CrossRef
31. Keren O, Yupatov S, Radai MM, et al. Heart rate variability (HRV) of patients with traumatic brain injury (TBI) during the post-insult sub-acute period. Brain Inj. 2005;19:605–611
    • View In Article
    • MEDLINE
    • CrossRef
32. Rappaport M, Hall KM, Hopkins K, Belleza T, Cope DN. Disability rating scale for severe head trauma: coma to community. Arch Phys Med Rehabil. 1982;63:118–123
    • View In Article
    • MEDLINE
33. Sacha J, Pluta W. Different methods of heart rate variability analysis reveal different correlations of heart rate variability spectrum with average heart rate. J Electrocardiol. 2005;38:47–53
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (389 KB)
    • CrossRef
34. Malik M. Heart rate variability: standards of measurement, physiological interpretation and clinical use (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology). Circulation. 1996;93:1043–1065
    • View In Article
    • MEDLINE
    • CrossRef
35. Bakeman R. Recommended effect size statistics for repeated measures designs. Behav Res Methods. 2005;37:379–384
    • View In Article
    • MEDLINE
    • CrossRef
36. Tabachnick BG, Fidell LS. Using multivariate statistics. Boston: Pearson Education; 2007;
    • View In Article
37. Bilotta F, Branca G, Lam A, Cuzzone V, Doronzio A, Rosa G. Endotracheal lidocaine in preventing endotracheal suctioning-induced changes in cerebral hemodynamics in patients with severe head trauma. Neurocrit Care. 2008;8:241–246
    • View In Article
    • CrossRef
38. Leone M, Albanèse J, Viviand X, et al. The effects of remifentanil on endotracheal suctioning-induced increases in intracranial pressure in head-injured patients. Anesth Analg. 2004;99:1193–1198
    • View In Article
    • MEDLINE
    • CrossRef
39. Baguley IJ. The excitatory:inhibitory model (EIR model): an integrative explanation of acute autonomic overactivity syndromes. Med Hypotheses. 2007;70:26–35
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (206 KB)
    • CrossRef
40. Blackman JA, Patrick PD, Buck ML, Rust RS. Paroxysmal autonomic instability with dystonia after brain injury [published erratum appears in Arch Neurol 2004;61:980]. Arch Neurol. 2004;61:321–328
    • View In Article
    • MEDLINE
    • CrossRef
41. Wang VY, Manley G. Recognition of paroxysmal autonomic instability with dystonia (PAID) in a patient with traumatic brain injury. J Trauma. 2008;64:500–502
    • View In Article

• a ADInstruments Pty Ltd, Unit 13, 22 Lexington Dr, Bella Vista, NSW 2153, Australia.
• b SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.